Browsing by Author "Chen, Hsuan-Sheng"

Now showing 1 - 7 of 7

A framework for video event classification by modeling temporal context of multimodal features using HMM
(2014-02-01)
Chen, Hsuan-Sheng
;
Tsai, Wen-Jiin
;
資訊工程學系
;
Department of Computer Science
Semantic high-level event recognition of videos is one of most interesting issues for multimedia searching and indexing. Since low-level features are semantically distinct from high-level events, a hierarchical video analysis framework is needed, i.e., using mid-level features to provide clear linkages between low-level audio-visual features and high-level semantics. Therefore, this paper presents a framework for video event classification using temporal context of mid-level interval-based multimodal features. In the framework, a co-occurrence symbol transformation method is proposed to explore full temporal relations among multiple modalities in probabilistic HMM event classification. The results of our experiments on baseball video event classification demonstrate the superiority of the proposed approach. (C) 2013 Elsevier Inc. All rights reserved.
A SINGLE-SCAN ALGORITHM FOR MINING SEQUENTIAL PATTERNS FROM DATA STREAMS
(2012-03-01)
Li, Hua-Fu
;
Ho, Chin-Chuan
;
Chen, Hsuan-Sheng
;
Lee, Suh-Yin
;
資訊工程學系
Sequential pattern mining (SPAM) is one of the most interesting research issues of data mining. In this paper, a new research problem of mining data streams for sequential patterns is defined. A data stream is an unbound sequence of data elements arriving at a rapid rate. Based on the characteristics of data streams, the problem complexity of mining data streams for sequential patterns is more difficult than that of mining sequential patterns from large static databases. Therefore, mining sequential patterns from data streams is a challenging research issue of data mining and knowledge discovery. Hence, an efficient single-pass algorithm, called IncSparn (Incremental Sequential pattern mining of streaming itemset-sequences), is proposed for discovering sequential patterns from streaming itemset-sequences over extended sliding window models. In the framework of IncSpam algorithm, a new sliding window model, called CSW-BV (Customer Sliding Window with Bit-Vectors), and an extended lexicographic tree-based data structure, called LesSeq-Tree (Lexicographic Sequence Tree), are developed to reduce the time and memory needed to slide the windows over streaming data and maintain all sequential patterns of current sliding windows. Experimental results show that the proposed method is an efficient single-pass algorithm for mining sequential patterns from streaming data.
A trajectory-based ball tracking framework with visual enrichment for broadcast baseball videos
(2008-01-01)
Chen, Hua-Tsung
;
Chen, Hsuan-Sheng
;
Hsiao, Ming-Ho
;
Tsai, Wen-Jiin
;
Lee, Suh-Yin
Pitching contents play the key role in the resultant victory or defeat in a baseball game. Utilizing the physical characteristic of ball motion, this paper presents a trajectory-based framework for automatic ball tracking and pitching evaluation in broadcast baseball videos. The task of ball detection and tracking in broadcast baseball videos is very challenging because in video frames, the noises may cause many ball-like objects, the ball size is small, and the ball may deform due to its high speed movement. To overcome these challenges, we first define a set of filters to prune most non-ball objects but retain the ball, even if it is deformed. In ball position prediction and trajectory extraction, we analyze the 2D distribution of ball candidates and exploit the characteristic that the ball trajectory presents in a near parabolic curve in video frames. Most of the non-qualified trajectories are pruned, which greatly improves the computational efficiency. The missed balls can also be recovered in the trajectory by applying the position prediction. The experiments of ball tracking on the testing sequences of JPB, MLB and CPBL captured from different TV channels show promising results. The ball tracking framework is able to extract the ball trajectory, superimposed on the video, and in near real-time provide visual enrichment before the next pitch coming up without specific cameras or equipments set up in the stadiums. It can also be utilized in strategy analysis and intelligence statistics for player training.
DISCOVERING EMERGING MELODY PATTERNS FROM CUSTOMER QUERY DATA STREAMS OF MUSIC SERVICE
(2011)
Li, Hua-Fu
;
Chen, Hsuan-Sheng
;
資訊工程學系
;
Department of Computer Science
Mining melody structure patterns from music query data is one of the most interesting issues of multimedia data mining. In this paper, we introduce a new kind of melody structure pattern, called emerging melody pattern (EMP), for knowledge discovery from music query data streams. EMPs are defined as music data items with melody item-sets whose support increase significantly from one sliding window to another window from streaming melody data sequences. An efficient data mining algorithm, called MEMPA (Mining Emerging Melody Pattern Algorithm), is proposed to discover all EMPs from music query data. In the framework of MEMPA, a prefix tree-based structure, called EMP-tree (Emerging Melody Pattern tree), is developed for storing EMPs effectively from current stream sliding windows. Experimental results show that the proposed algorithm is an efficient method for discovering all EMPs from streaming melody sequences.
HUMAN ACTION RECOGNITION BASED ON LAYERED-HMM
(2008)
Wu, Yen-Chieh
;
Chen, Hsuan-Sheng
;
Tsai, Wen-Jiin
;
Lee, Suh-Yin
;
Yu, Jen-Yu
We address the problem of human action understanding of the upper human body from video sequences. Time-sequential images expressing human actions are transformed to sequences of feature vectors containing the configuration of the human body. A human is modeled as a collection of body parts, linked in a kinematic structure. The relation of the joints is used to estimate the human pose. A proposed layered HMM framework decomposes the human action recognition problem into two layers. The first layer models the actions of two arms individually from low-level features. The second layer models the interrelationship of two arm as an action. Experiments with a set of six types of human actions demonstrate the effectiveness of our proposed scheme, and the comparisons with other HMM systems show the robustness.
Incorporating frequent pattern analysis into multimodal HMM event classification for baseball videos
(2016-05)
Chen, Hsuan-Sheng
;
Tsai, Wen-Jiin
;
資訊工程學系
;
Department of Computer Science
Data mining and frequent pattern analysis have recently become a popular way of discovering new knowledge from a data set. However, it is rarely applied to video semantic analysis. Therefore, this paper introduces two methods: frequent-pattern trained HMM and frequent-pattern tailored HMM to incorporate frequent pattern analysis into multimodal HMM event classification for baseball videos. Besides, different symbol coding methods including temporal sequence coding and co-occurrence symbol coding for multimodal HMM classification are compared. The results of our experiments on baseball video event classification demonstrate that integration of frequent pattern analysis could help to improve event classification performances.
Pitch-by-pitch extraction from single view baseball video sequences
(2007)
Chen, Hsuan-Sheng
;
Chen, Hua-Tsung
;
Tsai, Wen-Bin
;
Lee, Suh-Yin
;
Yu, Jen-Yu
This paper presents a novel method for reducing a baseball video segment from one batter to next batter into a more compact pitch-by-pitch video by pitching ball trajectory detection. The pitch-by-pitch video shows the complete pitching and batting process and largely reduces the source video data, making pitching analysis of broadcast baseball sequences an easier task. The proposed method has been tested for several long sequences, and promising results are reported.